ruby-crb-blast

Run conditional reciprocal best blast

CRB-BLAST is a novel method for finding orthologs between one set of sequences and another. This is particularly useful in genome and transcriptome annotation. CRB-BLAST initially performs a standard reciprocal best BLAST. It does this by performing BLAST alignments of query->target and target->query. Reciprocal best BLAST hits are those where the best match for any given query sequence in the query->target alignment is also the best hit of the match in the reverse (target->query) alignment. Reciprocal best BLAST is a very conservative way to assign orthologs. The main innovation in CRB-BLAST is to learn an appropriate e-value cutoff to apply to each pairwise alignment by taking into account the overall relatedness of the two datasets being compared. This is done by fitting a function to the distribution of alignment e-values over sequence lengths. The function provides the e-value cutoff for a sequence of given length.